The transparency imperative applies to the records of judicial proceedings just as fully as it does to the proceedings themselves. Access to records in criminal cases is essential because nearly all prosecutions today are resolved through plea deals, with no formal evidentiary proceedings.3 3. According to the U.S. Sentencing Commission, in FY2023 more than 97% of the sentences imposed in federal court were entered on a plea of guilty to the crime charged. U.S. Sentencing Commission, 2023 Annual Report at 16, https://www.ussc.gov/about/annual-report-2023The same is true in civil litigation, where federal data shows that trials are held in less than 1 percent of civil lawsuits, with most resolved through dismissal, summary judgment, or settlement.4 4. Jeffrey Smith and Grant McQueen, “Going, Going, But Not Quite Gone: Trials Continue to Decline in Federal and State Courts. Does it Matter?”, Judicature, Vol. 101, No. 4 (Winter 2017), https://judicature.duke.edu/articles/going-going-but-not-quite-gone-trials-continue-to-decline-in-federal-and-state-courts-does-it-matter/The public’s ability to follow and understand what transpires within the justice system is merely theoretical if the motion papers and other material filed with the court are not available for review. The common law has long acknowledged just this, recognizing an affirmative, qualified, public right to inspect and copy judicial records.5 5. See, e.g., Nixon v. Warner Commc’ns, Inc., 435 U.S. 589, 597 (1978); United States v. Amodeo, 71 F.3d 1044, 1048 (2d Cir. 1995); In re Cont’l Ill. Sec. Litig., 732 F.2d 1302, 1308 (7th Cir. 1984). Many courts—state and federal—have found that the First Amendment independently ensures public access to most judicial records.6 6. E.g., U.S. v. Ochoa-Vasquez, 428 F.3d 1015, 1030 (11th Cir. 2005); In re Search Warrant, 855 F.2d 569, 573 (8th Cir. 1988); In re NBC, Inc., 828 F.2d 340, 346-47 (6th Cir. 1987); In re N.Y. Times Co., 828 F.2d 110, 114 (2d Cir. 1987); Associated Press v. In re Washington Post Co., 807 F.2d 383, 390 (4th Cir. 1986); In re Globe Newspaper Co., 729 F.2d 47, 52 (1st Cir. 1984); Associated Press v. U.S. Dist. Ct., 705 F.2d 1143, 1145 (9th Cir. 1983); Grube v. Trader, 420 P.3d 343, 353 (Haw. 2018); Daily Press, Inc. v. Commonwealth, 739 S.E.2d 636, 640 (Va. 2013); Republican Co. v. Appeals Ct., 812 N.E.2d 887, 892 n. 8 (Mass. 2004); Skolnick v. Altheimer & Gray, 730 N.E.2d 4, 16 (Ill. 2000); Wichita Eagle Beacon Co. v. Owens, 27 P.3d 881, 883 (Kan. 2001); State v. Schaefer, 599 A.2d 337, 342 (Vt. 1991).

Yet, many state criminal court records, tax court records, and even records of certain federal civil proceedings are not available for inspection online. It is high time to end this impediment to meaningful judicial transparency and mandate online, electronic access to judicial records as the default rule. This can largely be done without new technology investments, without undermining privacy interests, and without imposing new administrative burdens on court clerks.

State Court Transparency Problems

At the dawn of the internet, when court records were filed on paper and transferred to digital form by the clerk’s office, many states adopted rules prohibiting public online access to criminal court records to protect private data and to avoid the dissemination of incorrectly transcribed information.7 7. See, e.g., Electronic Access to Criminal and Family Court Records: Hearing on S. 246 Before the U.S. Judiciary Comm., 2007-2008 Leg. (Vt. Jan.11, 2008) (audio recording , pt.2, at 10:03-14:16); Electronic Access to Criminal and Family Court Records: Hearing on S. 246 Before the S. Judiciary Comm., 2007-2008 Leg. (Vt. Jan.11, 2008)(audio recording, pt.3, at 30:09-31:31) (indicating that Vermont prohibited online criminal record access out of concern that dockets placed online could contain errors generated by court clerks transcribing information from paperfiles by hand, and to protect litigants’ ostensible privacy interests).But today, lawyers are nearly universally required to file motion papers electronically and have an ethical obligation to redact certain private data from judicial records. The reasons for the initial restrictions on remote electronic access no longer apply, yet many persist. Nearly two-thirds of states still have rules limiting public online access to criminal court records to some degree even though the need for online record access applies with particular force to criminal proceedings, where the public and press’s interest in real-time oversight is at its peak, and where criminal defendants, once charged, possess significantly diminished privacy interests in the unsealed records of their cases.

The access restrictions vary from state to state. Typically, the technology is in place to allow online access to the court file in a criminal prosecution, but remote access is restricted to the parties and their attorneys. The public is denied remote access to motion papers and related filings in criminal prosecutions, so these can be viewed only at the courthouse, in person, during operating hours, where the records are accessed electronically through an online terminal, just as they could readily be accessed from home or an office. These bans on remote access make it extremely burdensome for journalists and members of the public to follow the course of a criminal prosecution and to monitor the functioning of the criminal courts. Simply learning the reasoning behind the parties’ motions or assessing the reasonableness of their positions requires a significant investment of time and effort. And to obtain a copy of a record for future reference, a reporter must generally pay the clerk to print a hard copy or take pictures of each page on the court terminal with a phone.

Federal Court Transparency Problems

The same prohibition against public online access to motions and other filings readily available to the parties exists in federal tax courts. Tax Court Rule 27(b) draws a strict line between party access and public access, providing public online access only to the docket and any court opinion or order. Full electronic access to case files is only available in person at the courthouse.

Federal Rule of Civil Procedure 5.2(c) similarly prohibits online press or public access to, inter alia, any “action or proceeding relating to an order of removal, to relief from removal, or to immigration benefits or detention.” Under this rule, “[u]nless the court orders otherwise,” public online record access is restricted to the docket sheet and any court order ultimately entered; individuals must travel to the courthouse to access motion papers, supporting evidentiary submissions, and all other records through an online terminal. Because of Rule 5.2(c), more than 50,000 habeas corpus petitions filed in recent immigration cases are public only in theory.8 8. Madeleine O’Neill, “The Rabbit Hole: How the search for one court record became a nationwide movement.” Court Watch (June 7, 2026), https://www.courtwatch.news/p/the-rabbit-hole-how-the-search-for-one-court-record-became-a-nationwide-movement.

Each of these state and federal rules reverses the working presumption at common law that judicial records are not hidden from public view. They deny meaningful access to records the public has a substantial interest in seeing, make it more likely that misconduct and abuse will go undetected, and reduce public confidence that justice is being done.

A Straightforward Solution

Because most courts now require parties to submit records online in digital format, and maintain those records electronically, the public right of access must be recognized as a right to access those records where they exist, online. The internet has become the “principal source[] for knowing current events,”9 9. Packingham v. North Carolina, 582 U.S. 98, 107 (2017). and requiring a courthouse visit to inspect a record imposes a serious impediment to the transparency that is “pivotal to public perception of the judiciary’s legitimacy and independence.”10 10. U.S. v. Aref, 533 F.3d 72, 83 (2d Cir. 2008) (internal quotations and citation omitted).

This impediment can readily be removed by amending the offending court rules or, if needed, through legislation declaring a qualified right of online public access to court records. Such a law would:

• Contain legislative findings about the importance of transparency to maintain confidence in our system of justice and affirm the value of contemporaneous online public access to court records;
• Define broadly the “judicial records” presumptively available for online public inspection to include everything filed with a court that is relevant to the performance of the judicial function and useful in the judicial process, regardless of whether a judge ultimately chooses to use the record in determining a litigant’s substantive legal rights; and11

11. See Lugosch v. Pyramid Co. of Onondaga, 435 F.3d 110, 121-123 (2d Cir.) (explaining need for broad definition of the “judicial records” presumptively available for public inspection).

• Require that such judicial records must be available to the public online, absent a judicial determination, made on a record-by-record basis, that a compelling governmental interest requires an exemption from the requirement for online access.

The law would remove any requirement for physical presence in a courthouse to examine a judicial record but would not limit a court’s ability to seal records from public view under existing standards. Privacy, safety, and other considerations currently deemed sufficient to justify sealing a record or otherwise restricting public access would not be altered, but the practice of keeping large categories of court records behind a barrier would be ended.

Acknowledgements

This comment draws on research by Yale Law students Hannah Berkman, Abi Heath, Ara Johnson, Nick Jones, Bryce Morales, Ryan Fore, and Anna Selbrede.

The following can be attributed to Katie Fallow, deputy litigation director at the Knight First Amendment Institute at Columbia University:

“We welcome Judge Kelly’s order, but we are concerned that the White House continues to ban some news outlets from the White House press pool, where President Trump speaks the most with reporters. We are also concerned that the White House is still threatening to ban other journalists from the White House because he disagrees with their reporting. That practice is unconstitutional, and harms both the press and the public, which relies on journalists to report independently on the president and his administration.”

Judge Kelly ruled that the ban violated the news outlets’ Fifth Amendment due process rights and violated the First Amendment because the ban retaliated against the news outlets based on their speech, and because the ban was an unconstitutional viewpoint-based restriction in a nonpublic forum. The ruling did not address the White House’s continued ban on CNN and Politico serving in the White House press pool, the group of reporters who cover the president in smaller spaces like the Oval Office and aboard Air Force One.

In an earlier phase of the case, the Knight First Amendment Institute joined the Reporters Committee for Freedom of the Press (RCFP) and more than 75 media organizations in filing an amicus brief in support of the news organizations. 

Read Judge Kelly’s decision here.

Read more about the case, CNN v. Trump, here. 

The Knight Institute has supported the Associated Press in its separate challenge to its exclusion from the White House, filing three separate amicus briefs over a period of 18 months. The Institute has also supported The New York Times in its challenge to restrictions on access to the Pentagon. Earlier, the Institute successfully sued President Trump over his blocking of critics from his Twitter account, which the Institute argued was a “public forum” for First Amendment purposes.

For more information, contact: Lorraine Kenny, lorraine.kenny@knightcolumbia.org. 

With the start of a new academic year upon us, as approximately 150,000 students are in their first weeks of law school, I’m writing a series of reflections on the conversations we had with podcast guests, thinking about the future of legal education and of professional responsibility. Other posts in this series are available here.

***

The deals that several Big Law firms made with the Trump administration in the first months of 2025 came up repeatedly in our podcast conversations, as we contemplated their significance for the legal profession. Professor Lawrence Lessig, the guest for our second episode, offered his concept of “institutional corruption” to explain why these firms chose to make agreements that are so repugnant to the legal profession’s self-conception as well as its duty to be independent.

Lessig defines institutional corruption as influences, especially economic ones, that guide an institution away from what its underlying purpose is supposed to be. In our conversation, he applied it to these firms:

I think it’s because these people have begun to conceive of their job as the job of just maximizing the wealth of their firm… And that’s what we’re seeing in corporations today. You say you’re a news organization. No, no, you’re a corporation. Your job is to maximize wealth. So, if you’ve got this tyrant who comes along, and says, 'Sell your news department to me, and bend over backwards,' if that increases the shareholder value, that’s your job. You have no choice to do anything else except that. We increasingly see it in other organizations that also have kind of money at their core. So, we certainly see it in the context of these law firms who I think just thought, ‘This is simple. We can make this problem go away, and we’ll still have our bonuses at the end of the year….’ And if you don’t have people who are willing to say, ‘I get this is how I maximize my wealth, but this is how I am human. This is what I do to be human,’ we’re lost. We need those people, and we don’t have them right now. (9:38-12:53)

Our conversation with Lessig helped me pinpoint why the law firm agreements with Trump made me so angry at the time they were struck, and why I continue to think of them as such a significant event in the history of this administration’s authoritarian attacks. It was the scale of the trade these firms decided to make, and how easily they made it: protecting their profits outweighed even the imperative to resist an attack on the foundations of the rule of law itself. The leadership of those firms knew as well as any of us that if law firms are not free to represent any client of their choosing, the concept of the independent bar fails. Equally true, if these firms’ potential clients cannot trust that their lawyers are not looking over their shoulder to weigh whether a future administration might attack them for representing that client, or for particular arguments they make on their behalf, they will lose faith in these firms, and perhaps in the legal system itself.

These firms knew all of this, and made a deal anyway. They did it even though they have ample financial resources and legal firepower to fight back, to a far greater degree than some of the other independent institutions Trump has similarly attacked. It is simply shameful.

How do we hold them accountable for the choices they made? Although I believe Lessig’s institutional corruption theory is correct, it is hard to envision Big Law firms becoming less profit-driven. Nor is the decision to make a deal with the Trump administration, no matter how ill-advised, a matter for disciplinary measures from bar authorities.

However, another path to accountability lies in making sure that this episode is not forgotten by the public, or allowed to fade into obscurity. Democrats in Congress have already signaled that they will investigate the nine firms that made deals with the Trump administration, whether through hearings or written requests for information. The Knight Institute has filed a FOIA request asking for the text of these agreements, which have still not been made public.

The deals these firms made are a matter of enormous public interest: They go to the heart of public confidence in the integrity of the legal profession. The public has a right to know what these firms discussed with the administration, what they agreed to, and what justifications they can offer for their decisions. Law students deserve to know more about the character of the leadership at law firms they are considering working for. These firms’ present and potential future clients deserve to know whether their lawyers are willing and able to provide the “zealous advocacy” that legal ethics requires. And the firms themselves, along with other powerful institutions that chose to capitulate to unconstitutional pressure rather than to fight back, need to see that these choices will become an enduring part of their legacy.

In debate to override President Ford’s veto of the 1974 FOIA amendments, Rep. John Dent (D-PA) observed that, in July 1776, our Republic was founded by a people “tired of hearing nothing from Mother England but dictums as to how to conduct their affairs and where they were to send taxes.” Now, 250 Julys later, we come back to government by imperial decrees which rebuke explanation. OMB has explicitly divested the executive branch from transparency. In its own words:

The number of FOIA requests from leftist activist groups has exploded due to Trump Derangement Syndrome and it would not be serving the American peoples’ desire for a smaller, leaner government to waste money hiring additional FOIA processors to address this bourgeoning cottage industry of political activists. In working with limited budgets, we choose to hire ICE officers instead of FOIA processors.

Of course, blaming mentally deranged leftists for FOIA’s collapse is as unfounded as “Trump Derangement Syndrome” itself.1 1. A search of the FOIA Project database for new FOIA litigation since July 1, 2025, shows the America First Legal Foundation filed 11 FOIA lawsuits, while the right-leaning Judicial Watch filed 55. Impressive numbers! It is also (hopefully) insincere, except as an expression of political loyalty. It is also reflective of how awash we are in, to use a technical term, bullshit.

FOIA cannot seem to escape bullshit

FOIA cannot seem to escape bullshit. Supporters speak as though it provides meaningful on-demand transparency, because we believe it is worth fighting for and possible to achieve. Others find FOIA a convenient prop. FBI Director Kash Patel accused the Atlantic magazine of purposefully failing to file “a simple FOIA” before reporting that the FBI requested breaching equipment to open his hotel room when he was nonresponsive. Perhaps the Atlantic did not turn to FOIA because another request seeking similar materials was given an estimated response date of 2031. Prolific FOIA user and litigant Robert F. Kennedy Jr. became Health and Human Services Secretary on a promise of “radical transparency” and, in his first hundred days, eliminated the entire Centers for Disease Control FOIA office and most of the FOIA staff at the Food and Drug Administration and the National Institutes of Health.

Maybe FOIA has never worked. The concerns of 1974 are essentially unchanged, excepting that the conferees of the 93rd Congress had yet to learn of the Deliberative Process Privilege.2 2. Access to agency deliberations to learn real reasons for agency action, and not just the polished final published action, was an explicit purpose of Congress. As Sen. Ted Kennedy, Senate sponsor of the 1974 amendments, explained in his introduction of the bill on May 30, 1974: “The people can judge public officials better by knowing what they are doing, rather than only by listening to what they say. But to know what Government officials are doing, the people must have access to their decisions, their orders, their instructions, their deliberations, their meetings.” Fifty-two years ago, Congress knew that “if there is a way that a law can be interpreted to promote secrecy and to deny the public access to public records, some Government officials will find that way.”3 3. Rep. Thone (R-NE), 120 Cong. Rec. H1787-H1803 at 251 (1974). FOIA was strengthened in response to the realization that, faced with an “imperial Presidency . . . . and the inherent rules and criteria-making powers of the bureaus and departments of the executive branch, Congress ha[d] become the fifth wheel on a hearse.” As Rep. Bill Alexander (D-AR) observed in urging the override of President Ford’s veto of the 1974 bill, as patriotism is the last refuge of scoundrels, so too “withholding of information from the public is the ‘last refuge’ of the bureaucrat.”

FOIA is great because it tasks agencies with opening their books to public inspection. This is also its fundamental weakness: it is administered by agencies, which are staffed by people. Those people, even where decent, public-minded, and conscientious, are dedicated to their agency’s mission. Even worse, some of those people are lawyers, specifically trained to protect their clients’ interests at all costs, and those interests do not align with disclosing agency secrets. The adversarial American legal system, the legalistic nature of government, and the threat perceived from FOIA make on-demand disclosure a fantasy. FOIA is, at best, an inconvenient unfunded mandate, if not a threat to agency mission or, even worse, a career. Some research suggests merely receiving a request can harm civil servants’ trust in the public they serve.

FOIA Processing

As a litigator, and until recently an agency FOIA processor, I will focus on what I know best: processing and exemptions. But what is processing? And why is it so difficult and expensive? Releasing a record is the easiest thing in the world. Simply attach it to an email. Hell, run off a few hundred copies and throw them off the roof, you silly goose. Processing is not about releasing records though; and it is slow and expensive by design. It is a painstaking series of redundancies taken by risk-averse FOIA offices to protect themselves from internal institutional backlash:

1. Find the record
2. Identify the relevant interests
   1. is the requester interested in this record, i.e., is it responsive?
   2. will release harm an interest the agency cares about?
      1. inquire with program office
      2. inquire with interested agency leadership
      3. is the agency legally allowed to care about that interest?
         1. if not, is there another plausible interest that can be asserted for the same info?
   3. is the agency legally required to care about any other interest?
   4. is someone else interested?
      1. refer for interagency consultation
      2. ask private partners if they would like it kept confidential
   5.  is there an agency policy or custom to always redact certain kinds of information regardless of interest?
3. Redact information affecting all identified interests
4. Repeat step 2 & 3 twice each with your supervisor and manager
5. Awareness Review!
   1. repeat step 2 & 3 with interested program offices
   2. repeat step 2 & 3 with interested leadership office
   3. make appropriate changes
   4. repeat Step 4 to address changes made in the awareness consultation

“Processing” is designed to explore every plausible alternative to release. If someone in that chain refuses or forgets to clear the release it can go on the shelf indefinitely. The process invites stakeholders from any interested office to substitute their judgment for the FOIA expert’s, with formal accountability mechanisms few and far between.4 4. Some particularly clever administrators even realized that nonresponsive documents do not appear in Vaughn indices or determination letters, making them de facto unreviewable. Can’t argue a plausible sounding exemption? Just re-run the search! If inexperienced agency leaders take the wheel and crash into litigation, the responsible FOIA office is the moral crumple zone.

Nine narrowly tailored exemptions have thus metastasized into a “veritable avalanche of FOIA related precedent” as agencies go to “great expense of litigation . . . display[ing] an attitude that could be interpreted as . . . ‘If you want this information, sue the Government’” before releasing documents under protest. From personal experience, the presumption of disclosure is a dead letter. Recommended releases are pared down by every step, and even a combative pro-disclosure FOIA officer must pick their battles carefully. More career-minded experts quickly learn management’s preferences for what kinds of information to withhold, and personal peccadilloes quickly accrete into permanent presumptions. Virtually every email address in government is withheld as a “clearly unreasonable invasion of personal privacy” under Exemptions 6, 7(C), and (at DHS) even 7(F)! (Exemption 7(F) applies where release of the information would endanger the life of a “small and specific group,” such as the “family members or coworkers of a named individual” and is not generally used for stuff like desk phone numbers and emails.)

The broad nature of the rules thus facilitate abuse, and past legislative fixes, like the foreseeable harm standard from the 2016 FOIA Improvement Act, have fallen short. It has been leveraged in litigation, but for most retail requesters, the only change is to the boilerplate: “Please know I have considered the foreseeable harm standard in making this determination.” Requesters want to be taken seriously. They turn to litigation for the same reason we jam “0” and shout “agent” at customer service phone trees: the process seems designed to frustrate you into giving up.

Taking requesters seriously would break administrative kayfabe: the FOIA office can’t admit they do not actually make the FOIA decisions. FOIA decisions come from agency managers, leadership, and attorneys who believe their role, understandably, is to “protect the agency.” And these days, agency managers, leadership, and attorneys aren’t acting their best. The job of the FOIA office is thus not to release records, but to generate plausible-sounding reasons for requesters and courts to believe they simply couldn’t be released. The denial is intended to persuade without regard for truth, often simply by reciting the statutory standard and claiming it applies. It is exactly Harry Frankfurt’s definition of bullshit, and it is the legacy of FOIA. President Ford even vetoed the 1974 amendments for bullshit reasons:

To my agency readers, I know very well this is not true of all agencies, FOIA offices, or especially FOIA professionals. When I say “agencies,” I’m talking about those other bad ones, not you. Most of us are doing the best we can with what we have. But laws cannot be written assuming good faith compliance.

So What?

What, then, is to be done? First, and most importantly, actually fund FOIA’s mandate. Give every FOIA office a billion dollars, sure. Second, make sure they actually use those resources on transparency, not massaging the secrecy preferences of agency leadership, by getting the extra cooks out of the kitchen.

At least for appeals, prohibit anyone involved in denying a request from being part of the appeal. I heard too many orders to deny a request on shaky grounds because “we can sort it out if they appeal” and saw appeals decided by the same staff and managers responsible for the denial (none of whose names, incidentally, appear on the release letter). Ideally, appeals would be decided by an independent body unbeholden to the agency’s parochial interests. Reps. Frank J. Horton (R-NY) and John Erlenborn’s (R-IL) proposed FOIA amendments in H.R. 4960 included the creation of an FOI Commission to resolve appeals from agency denials. Rep. Erlenborn prophesized that the coming years would demonstrate the utility of such an idea. Prof. Margaret Kwoka’s proposal for an American Information Commission carries this argument forward far better than I could. Do whatever she says there and give it a billion dollars as well.

Determination letters are among the greatest sources of FOIA bullshit, as form letters written by nobody and signed pro forma. The 93rd Congress wisely realized that bureaucrats are less likely to behave arbitrarily when “the names and positions of all government officials responsible for [FOIA denials] are required . . . to be noted in denials and reported annually to Congress.” They did not anticipate that agencies would respond by simply hiding the ball. I wrote dozens of determination letters both releasing and withholding tens of thousands of pages of records which I personally reviewed, and my name never appeared in one. Initial and appellate determinations should name anyone who even looks at a proposed response. If meddling in FOIA can’t be banned in a way that matters, it should at least require accountability. Determinations should further list the filenames or titles of any fully denied record so that requesters can know what exactly, they are even appealing.

For appeals to be effective, a requester also has to know what they are appealing and why. The designer of the checkbox form letter which lists every existing exemption that may or may not apply should be ejected from the federal service. Determinations should include a plain list of each responsive withheld record by filename and the exemption used to withhold each. Yes, this will be burdensome, but the logic of Vaughn and its progeny applies just as much to appeals as litigation. And as the D.C. Circuit explained in Mead Data Central, Inc. v. U. S. Dep't of the Air Force, 566 F.2d 242 (D.C. Cir. 1977), there is an easy alternative:

Certainly these procedures add significantly to the resource costs an agency must bear if it chooses not to disclosure [sic] material it has in good faith decided is exempt. Those burdens may be avoided at the option of the agency, however, by immediate disclosure. Congress has encouraged the agencies to disclose exempt material for which there is no compelling reason for withholding, and an agency's own balancing of the resource costs of justifying nondisclosure against the value of secrecy may provide a rough estimate of how compelling is its reason for withholding.

The interagency and intra-agency consultation process, which has developed an entire extra-legal administrative ecosystem all its own, must be subject to legislative control. Agencies should absolutely be prohibited from salami-slicing requests into their constituent atoms and assigning each to a separate processing track. Earlier this year, I reviewed a simple request for the DOJ Headquarters visitor log, maintained as a single excel spreadsheet by the Justice Management Division. JMD split the request and the record 25 ways, with 25 records, 25 processing numbers, one for each component office. Of these, 1 was responded to in full, 3 were acknowledged, 1 was delayed because the component demanded processing fees, 1 was denied in full, 12 were responded to in part, and 7 were ignored. The partial denials incorporated four different exemptions depending on the local custom of the component FOIA offices. This is madness, generates massive policy inconsistency, was borderline unappealable, and makes a mockery of the DOJ’s self-reported FOIA processing numbers. One request becomes 25 by administrative sleight-of-hand.

The judicial review provisions must be rewritten and strengthened. Decades of case law have taken FOIA far from its original intent and language to facilitate agency delay and carelessness. Agencies can and do change justifications for withholding in initial determinations, appeals, and in court. The agencies have sole access to the relevant evidence, and without extraordinary insider access or reporting, it is nearly impossible to overcome the presumption of good faith afforded their declarations. An agency which fails to justify withholdings or redactions in sworn declarations is often given two or more extra tries before a court will order in camera inspection or the records’ release. Thankfully, this trend may be reversing in some cases as judges tire of “the notion that agencies may repeatedly seek summary judgment until they eventually justify withholding relevant material.” (See Ryan Mulvey’s piece for the true cost of the judiciary’s lax approach to “de novo” review for litigants.) 

While there are times when a court should seek more information before ordering release, agencies have grown accustomed to the courts as soft play areas. Declarations and indices of withholdings are not taken seriously until they are not only in litigation, but after a first or sometimes second loss at summary judgment. Even RCFP v FBI, which famously put teeth into the foreseeable harm standard, invented a new rule for records whose “context and purpose . . . make the foreseeability of harm manifest” 3 F.4th 350, 372 (D.C. Cir. 2020). This “context and purpose” standard is a pressure valve allowing judges to deny access to the really important records even where the agency cannot be bothered to produce a legitimate statement of foreseeable harm. It is little wonder that bad habits from decades of lax FOIA practice bled their way into the clumsy release of material under the Epstein Files Transparency Act.

On the substance of the exemptions, virtually all of the records cited as FOIA success stories in 1974 would be withheld under modern practice. The direction of Internal Revenue Service investigations to target political enemies? No, that’s deliberative, privileged, also likely exempt under one or more parts of (b)(7), and you should feel silly for asking. FDA facility inspection reports and Social Security Administration medical survey reports on nursing homes receiving federal payments? Maybe you can get a summary with details withheld as confidential commercial information. Statements of DOJ intent on proposed mergers? Definitely privileged. FBI counterintelligence program guidelines would never get past (b)(7)(E), and Federal Housing Administration appraisal reports I do not believe are even considered agency records. Some records must remain secret, and any written standards for secrecy are invariably subject to abuse. The subjects people most critically need information about are inevitably those which the government has erected the highest walls around. Dragnet surveillance programs, the evidence and reasoning underlying health and food safety decisions, who profiteers from the President’s racial purges, how agency social media accounts were suborned to the libidinal urges of not-so-closeted internet racists, threats to federal science, and political interference with agency process are not things you are allowed to learn about via FOIA in 2026. Especially if you can’t afford going to court.

As Rep. Spark Matsunaga (D-HI) explained introducing H.R. 12471 on the House floor, “the scope of this legitimate shield . . . could be stretched to suit particular partisan or personal purposes. It could be extended to veil matters unfavorable to the custodian agency or embarrassing to the officials therein.”

The specific revisions to the text of § 552(b) are less important, for now, than the appreciation that edits are needed. For a start, Congress should define the privileges protected under Exemption 5 and their limitations. The great black maw of the deliberative process privilege should be carefully fenced off. Clear definitions should be written for terms like “policy decision,” “predecisional,” and “deliberative.” Someone should really figure out if the “chilling effect” averred by so many agency declarants is real, if agency program staff are even aware of what is and is not released, or if they respond rationally to FOIA disclosure at all. If the words “deliberative process” appear in an employee’s email signature, their records should be presumed releasable on punitive grounds.

Exemption 2 could be cut entirely without anyone noticing, and while we are at it, the handful of cases on Exemption 9 (wells) have all indicated it would benefit from greater clarity as to what it protects and why. Exemption 3 is a problem of dozens of unrelated statutes passed by Congresses grown increasingly casual about categorical exemptions from transparency. At minimum there should be clear rules for judges to examine agency determinations that an Exemption 3 statute applies at all, and each should be subject to a public interest balancing test.

Exemption 4 is the corporate version of the “withhold it because you want to” rule. Congress should undo the damage from Argus Leader, require a real showing by submitters of allegedly commercial information that release would be harmful, and weigh the public interest against that determination. The 92nd Congress believed that “deletion of formulas or statistics or figures may in many cases entirely fulfill the purpose of the fourth exemption.” They did not think that consultant reports constituting the entire factual record for decisions with real operative legal effect could be protected because “the way we format our slides” is a trade secret.

Exemption 6 should do away with the limitation to personnel, medical, and “similar files,” which has similarly been interpreted away to near meaninglessness, and provide clearer definitions of whose privacy is protected, when, and why. The slow creep of Exemption 6 has otherwise grown like a kudzu vine across all cell phone numbers, then all phone numbers, then all meeting locations and addresses, then all email addresses, and now at some agencies, even the names of any “lower level” employee. 

Exemption 7’s original application, decried in the 70s as overbroad and unprincipled, was only to investigative case files themselves, and not the broad swath of law-enforcement-adjacent material covered under the modern “colorable claim of a rational nexus” tests. Exemption 7(C) was written expressly to cover the subjects of investigations, not every visitor to DOJ HQ; and 7(E) was initially designed to cover “only specialized techniques not generally used in investigations,” not to protect the locations of every camera, scope of every questionable digital surveillance dragnet, or mere existence of every Palantir enforcement tool in the federal inventory. 

These questions will require careful consideration, and input from the people who practice FOIA, not just senior agency leaders and attorneys. What is important now is to recognize the need and form the kind of special congressional committees whose investigations led to the 1974 amendments. Many are skeptical that this kind of cooperation can ever happen again, but if ever there were a moment for public records, we are living in it.

The following can be attributed to Scott Wilkens, senior counsel at the Knight First Amendment Institute:

“President Trump and his former co-defendants should not be allowed to delay this case further. This report concerns allegations of serious misconduct by the nation’s highest public official, and the public has an extraordinary interest in seeing it. The court should reject this latest attempt at delay and hear the case on November 2.”

The case concerns Volume II of Special Counsel Jack Smith’s report, which details his investigation into the defendants—President Trump, his personal aide Walt Nauta, and Mar-a-Lago property manager Carlos De Oliveira—over the alleged mishandling of classified documents at Mar-a-Lago. In February 2025, the Knight First Amendment Institute and American Oversight separately sought access to the report, arguing that the public has a right to see it. After months of litigation—and a ruling by the U.S. Court of Appeals for the Eleventh Circuit that there had been “undue delay” in the district court proceedings—Judge Aileen Cannon permanently blocked the report’s release in February 2026 and denied the Knight Institute’s and American Oversight’s motions to intervene. The Knight Institute and American Oversight appealed. On September 30, the Eleventh Circuit set oral argument for November 2, after the appeal had appeared on several tentative argument calendars. In papers filed today, the Knight Institute and American Oversight argue that there is no justification for delaying the argument again. 

Read today’s brief in United States v. Trump here.

Detailed information on the case, including a timeline of significant court filings, is available here.

Lawyers on the case include Scott Wilkens, Jameel Jaffer, and Alex Abdo for the Knight First Amendment Institute, and David Buckner for Buckner & Miles.

For more information, contact: Lorraine Kenny, lorraine.kenny@knightcolumbia.org. 

In Red Lion Broadcasting Co. v. FCC, 395 U.S. 367 (1969), the Supreme Court upheld the “fairness doctrine,” which required every broadcaster “to give adequate coverage to public issues” that is “fair in that it accurately reflects the opposing views.” The doctrine was enforced primarily through the Commission’s licensing authority. The Court, acknowledging that the doctrine granted the FCC extraordinary authority to regulate speech and the press, held that such power was appropriate because radio and television frequencies are inherently scarce and the government thus had an obligation to regulate the content of broadcasts to ensure fair and balanced coverage on issues of public importance.

Unbeknownst to the justices, however, the Red Lion case originated in a “politically motivated campaign to use the fairness doctrine to harass stations airing right-wing commentary, an effort inspired and managed by the White House and the Democratic National Committee and largely financed with political contributions.” The Nixon administration and its allies then used the same playbook to lean on broadcasters perceived as unfavorable to the President. The FCC abandoned the doctrine during the Reagan administration, partly in recognition of the danger such regulatory authority posed to press freedom.

Although the fairness doctrine is now defunct, it underscores how the FCC’s powers to regulate broadcast licensees can be used to suppress viewpoints perceived as unfavorable to the party in power.

Much has changed since the Court decided Red Lion. The audience for over-the-air broadcast is shrinking while internet-based media like streaming grow in both viewers and listeners as well as influence on issues of public concern. (Simultaneously, broadcast remains an essential medium for huge numbers of Americans, especially those of fewer means, making government interference in content decisions that much more fraught.)

The conditions that previously supported the Supreme Court’s decision in Red Lion have thus completely changed. The Government, particularly an agency with potent licensing and merger review authority, should have no power to regulate broadcast content beyond any that the government currently possesses for any other expressive medium (namely, the power to regulate unprotected speech or restrictions that meet strict scrutiny, which few will). Congress should clarify as much in statute, or a challenge should be brought to the Supreme Court to overrule Red Lion.

Indeed, Red Lion has never been disavowed, although the justices have expressly declined to extend it to other, later-developed communications media, including cable television (Turner Broadcasting v. FCC) and the Internet (Reno v. ACLU), to which the “scarcity” rationale for regulation is plainly inapplicable.

As long as the Red Lion precedent stands, every president will have the authority to assert control over broadcast content. This threat has only been magnified under the current administration.

In response to congressional questioning regarding whether the FCC might take enforcement actions based on constitutionally-protected editorial decisions by media companies, Chairman Brendan Carr stated that the FCC has “walked away from enforcing the public interest standard, and I don't think that's a good thing.” He also cited the news distortion policy and the broadcast hoax rule, which derive from the public interest standard, and stated that his position “is that we should be enforcing those rules and policies.” And, in January 2026, the FCC issued guidance suggesting that it would narrow the “bona fide news” exception to the equal time rule, which permits candidates whose opponents are afforded access to a licensee’s facilities to demand an equal opportunity for airtime.

This May, the FCC issued a Public Notice to “remind broadcasters of their longstanding public interest obligations and further ensure that broadcasters are continuing to comply with the public interest obligations that underpin their licenses.” The notice opened with a reiteration of the spectrum scarcity rationale that grounded the Red Lion opinion. It further stated that although the FCC has reduced many burdensome broadcast regulations, “the role of the Commission in determining whether broadcasters have met their public interest obligations has not faded.” And, emphasizing the Commission’s public interest authority, the notice asserts, “Where the Commission finds that a broadcaster has failed to serve the public interest, the Commission may take appropriate action, including enforcement action, grant of a renewal application with conditions and/or on a short term basis, requiring a licensee to file an early license renewal application, or designating an application for hearing.”

Given the danger of direct censorship inherent in the misuse of the public interest standard, Congress should exercise its responsibility to preserve the free flow of newsworthy information to the public by clarifying the scope of the standard to prevent overreach by the Commission.

Another approach could be a legal challenge seeking to overturn Red Lion. An opportunity for such a case may be on the horizon, as the same day the FCC issued its May Public Notice, ABC filed early license renewal applications for its eight owned-and-operated stations pursuant to an April 28, 2026 FCC order. ABC included with each application a detailed public interest statement and an objection that describes the FCC’s early renewal mandate as an “effort to suppress free speech under the guise of bureaucratic process.”

However the law changes, Chairman Carr’s cloaking of his efforts to influence broadcast coverage in the public interest standard make it clear that it must change. Indeed, fewer things could be more outside the public interest than an FCC interfering in what we see and hear over the air.

What all our names collectively convey is that we are organizations of journalists that exist to defend the rights of our colleagues—that is the “press” in press freedom.

I have often been intrigued by the name of the leading press freedom group in Canada, which is called Canadian Journalists for Free Expression (CJFE). Consider for a minute what that name communicates.

In the U.S., the framing and public perception is that journalists and media institutions come together when their own rights are threatened. Their mandates are grounded in a conception of press freedom necessitating endless debates about who is a journalist, where the profession begins, and where it ends.

CJFE meanwhile says it defends “free expression and those with the courage to seek and report the truth.” Its mandate is more flexible. Its historic mission is grounded in the broader defense of speech rights, encompassing Canadians who express themselves in a variety of ways.

In the U.S., the subtext is, “We as journalists have come together to defend our rights and we invite the public to support us in this endeavor, because everyone benefits from the work that journalists do and when information circulates freely.”

In Canada, the subtext is, “The rights that guarantee the free exercise of journalism are the same ones that protect speech rights exercised by all Canadians—so when your rights are threatened, our rights are threatened. We as journalists stand with you.”

Recently, I’ve begun to wonder if that broader framework might make sense in the U.S. Instead of asking the U.S. public to stand with journalists when our rights are threatened, perhaps we as journalists and media institutions should find ways to stand with the public. This approach may have particular resonance in the U.S. at a time when the media environment is evolving rapidly and the line between journalism and other forms of expressive activity are becoming ever more blurred.

Flipping the script in this way could strengthen trust and engage the public at a time of deepening polarization. The traditional journalist’s refrain—“we are your eyes and ears” or “we are your surrogate and represent the public in places to which you don’t have access”—is less likely to resonate than “we stand with you in defending your rights.”

Second, it would better reflect the way journalism is increasingly exercised. In fact, it is often members of the public who serve as the eyes and ears of the media, providing firsthand documentation that journalists then verify, contextualize, and disseminate to a wider audience. This was apparent during Operation Metro Surge in Minneapolis in January, when reporting on police abuse, including the killing of two protesters, was based largely on cell phone footage uploaded by bystanders. While journalists have a distinct professional identity and social function, the legal framework that protects their rights while covering protests is the same one that protects the protesters themselves.

Third, this broader framing would reflect the legal reality of how journalists actually defend their rights when they go to court. For the most part, journalists do not have and do not claim any special rights. When media organizations engage in litigation to defend their rights, they usually argue that the free speech principles that protect the rights to all Americans must apply to journalists as well.

Of course, the counterarguments are also compelling. Broadening the mandate in the Canadian style could undermine the media’s perceived independence and drag it into unwanted controversies. It could erode professional solidarity and weaken funding structures, since press freedom organizations in the U.S. are often backed by the media industry and industry-aligned foundations. There are myriad practical questions. What kind of free speech cases might media and journalistic organizations take up? Who would raise their voice, the news organizations themselves or the press freedom groups that they support? What kind of interventions are appropriate? Public statements? Amicus briefs? Public policy advocacy? In fact, CJFE’s own experience is an object lesson in what can go wrong. The organization has not fully recovered from its near collapse in 2018 after the staff issued a statement condemning Israeli violence against journalists and protesters in Gaza.

But times are very different today, and so is the practice of journalism. By maintaining a strictly domestic focus—standing up for Americans exercising their First Amendment Rights—journalists and the institutions that represent them can help shift the public debate in this country away from the institutional role of the media and more toward the exercise of our collective rights. In terms of issues, media organizations might consider supporting legal efforts to defend the rights of all Americans to engage in protest, including the right to record police activity. That opportunity was brought home by a recent decision in the Second U.S. Circuit Court of Appeals brought by a non-journalist and involving activity that was not fundamentally journalistic. Media organizations could also consider the possibility of joining efforts to defend the speech rights of non-citizens which are constitutionally protected and under assault.

A broader defense of speech rights by U.S. media institutions would recognize that the distinction between journalism and other forms of expressive activity are becoming more difficult to parse. It would align more closely with the legal reality, which is that journalists do not enjoy “special” rights and do not generally seek them. And it would reflect the obvious truth that, when it comes to freedom of expression, we are all in this together.

“The court’s decision is disappointing, but we intend to appeal. Spyware manufacturers that participate in the persecution of journalists shouldn’t be able to operate with impunity, and U.S. courts must ensure that they are held accountable for their actions where those actions violate U.S. law,” said Carrie DeCell, senior staff attorney and legislative advisor at the Knight First Amendment Institute. “We will continue to seek justice on behalf of El Faro journalists in U.S. court, where the case belongs.”

In November 2022, the Knight Institute filed the lawsuit alleging that NSO Group violated U.S. law by developing, selling, and assisting in the deployment of Pegasus against El Faro. Between June 2020 and November 2021, El Faro’s employees were subjected to at least 226 Pegasus infections. Through these attacks, their iPhones were accessed remotely and surreptitiously, their communications and activities were monitored, and their personal data was accessed and stolen. In recent years, NSO Group’s spyware has been used by authoritarian and rights-abusing regimes to target journalists, human rights activists, and political dissidents around the world. This case is the first filed by journalists against NSO Group in U.S. court.

Read today’s decision here. 

Read more about the case, Dada v. NSO Group, here.

In addition to DeCell, lawyers on the case include Jameel Jaffer, Alex Abdo, Stephanie Krent, and Nicola Morrow for the Knight First Amendment Institute, and Corey Stoughton for Selendy Gay PLLC.

As is true in most other areas of life in 2026, AI is being floated as the solution to FOIA’s woes. The federal government is already bullish on it—the Department of Homeland Security, the National Archives, State, and Justice, among others, have dabbled in its use. These experiments can be understood as a response to both short- and long-term trends, from a depleted federal workforce to the ever-increasing number of requests submitted across the federal government. A record 1.7 million requests were submitted last fiscal year, a million more than a decade ago. And this all comes as the number of federal records being created and stored is exploding. Even if agencies’ FOIA resources had remained constant into the current administration (which they didn’t), the government would be falling behind.

The allure of AI to address these trends is understandable. As researchers like Jason Baron have noted, the ever-increasing volume of federal records makes AI-assisted searching a particularly attractive idea. Given how fast AI is progressing, though, it’s not hard to imagine it handling the entire FOIA pipeline: interpreting requests, locating records, reviewing and redacting them for exemptions, and responding to the requester. Deployed at scale, this technology holds out the promise of reviewing millions of pages of text, understanding decades of law, and tracking factual minutiae all without complaining about being sent to Siberia. In this vision of the future, AI is a wholesale replacement of the administrative process, cutting through the trends that had seemingly doomed FOIA. Assuming such a vision is technically possible, though, is it something we want to work towards?

Something clearly needs to change. After more than a decade of litigating requests for journalists and news outlets, I am acutely aware of FOIA’s problems and their effect on newsgathering. And after talking with countless government employees at mediations, conferences, and committees over the years, I understand many of the complexities within the government that have contributed to FOIA’s delays. But before introducing a radical new technology we must be clear-eyed about what we are trying to achieve, how we will mitigate risk, and the red lines that cannot be crossed. In short, we need a FOIA AI framework. I have endeavored to sketch one out here.

My proposal has two parts:

• First, we must assess the risk of introducing AI into any particular part of FOIA. The European Union’s AI Act provides a good basic formulation: what is “the probability of an occurrence of harm and the severity of that harm”? Risks should be categorized, and the most severe ones mitigated.
• Second, we must empirically test whether the core values of FOIA are being promoted by a particular implementation of AI. In a nutshell, those values are disclosing as much non-exempt information to the public as quickly as possible. We could test, for example, whether introducing an AI-powered search tool within an agency actually leads to faster request response times.

I want to focus on the first part of the framework in this piece, because the line from this White House has generally been an all-gas, no-brakes approach. We are charging into widespread AI adoption even though, as described below, there is serious risk to FOIA and not much theorizing about the nexus. In 2024, a subcommittee of the FOIA Advisory Committee (on which I served) called for the Department of Justice to issue guidance to agencies on their use of AI in the FOIA process. To date, that has not happened. And the Office of Management and Budget’s 2025 memo on AI, which calls for additional steps when it comes to “high-impact AI” within the government, does not include FOIA processing among its presumptive categories. That is a mistake that should be rectified.

A core part of risk management is evaluating the severity of harm that might result from any particular course of action. When it comes to FOIA, we should break out different parts of the administrative process for evaluation: on one end of the spectrum there are many basic, repetitive tasks where the potential harm is low. This includes actions like writing to the requester acknowledging receipt of their request, responding to basic queries from the requester (where is the request in the process, what is the estimated date of completion), and document management (like sorting and de-duplication). I classify these as low risk both because AI is pretty good at them and because the consequences of getting it wrong are not very serious.

At the other end of the spectrum are decisions at the core of FOIA: how to search for records, what to release, and what to withhold. The worst-case “harm” of getting it wrong here is the requester failing to get what they asked for. That could be disastrous for persons in immigration proceedings, who must generally FOIA their own files from the government. For the press and civil society, it’s going to be the deprivation of records promoting accountability and informed self-governance. These scenarios are unquestionably “high-impact”: FOIA has helped expose COINTELPRO, the torture of prisoners by the US government after September 11, the details of the Federal Reserve’s secret liquidity program, the Afghanistan Papers, the toll of civilian casualties from US-led airstrikes in the Middle East, the Robert Mueller investigation records, and so many more stories. These records have had profound implications in both domestic and foreign affairs, and that type of accountability must be preserved.

What is the likelihood of these most severe harms if AI were deployed to answer FOIA requests? So much depends on the technical details of the deployment, which are impossible to predict ahead of time. There are also endemic problems with AI, like its penchant for hallucinations, and those could certainly create problems in processing a request. But it is not hallucinations that worry me the most. It is the fact that AI can appear to be neutral and objective while it actually obscures and magnifies prejudices to a greater extent than other technologies. Each model contains biases from its training corpus, its fine-tuning, its background instructions, its prompts, and so on. Those reflect human biases stemming from deliberate decisions, predispositions, and prejudices. And they will, in turn, be imported into whatever system AI is inserted into.

Consider the State Department’s machine-learning declassification trial program. The system was not trained to produce an objectively “correct” result, but rather to match the decisions of the agency’s human reviewers. The same is true of an AI tool Health and Human Services has deployed, called “FRED,” that purports to “analyze, identify, and generate predictions of text for redaction” under FOIA. FRED was trained on certain inspection “forms before redaction and versions of those forms after redaction by FDA staff,” just as at State. The government might argue that AI matching the results of the human review is the correct result, but we know that FOIA officers routinely overredact records when reviewing them for release. If AI bots are trained against the output of human reviewers, similar over-redaction tendencies will emerge.

We should also consider that more insidious directions could be built into these systems. The current administration, for example, instructed FBI agents to “flag” any documents in the Epstein files that mentioned the president. Secretary Clinton set up a private email server, establishing a whole category of records that was never searched. ICE recently put out a purported policy of only releasing bodycam video when it is in the agency’s “best interests.” And so on. A bot that is trained to redact politically inconvenient information at scale, or instructed to never search a particular record repository, would pose grave risks to FOIA’s transparency mandate.

If FOIA is already experiencing these types of issues, though, perhaps a fairer measurement of risk is the delta between a human-centric FOIA system and one powered by AI. Even here, my sense is that it will be significant. Right now, line-level FOIA officers make the initial decisions about how to search for records and what to redact. Those decisions are subject to review and sometimes are overruled by higher-ups, to be sure. But in most cases the basic contours of disclosure are set by persons who have some distance from the White House, are trained to follow the law, and can push back against political meddling. A human reviewer might balk at a blatantly political instruction or blow the whistle about it, as happened with the Epstein files. Or humans might simply release the information anyway, as the Office of the Director of National Intelligence did last year with a memo on Venezuela.

Aggressively deploying AI could invert this structure, giving the White House near-complete control in FOIA responses from the get-go. If a FOIA bot is deployed to make the first cut on redactions, it could be given instructions like: redact anything that contradicts the president’s statements. AI can be trained from the outset to accept these types of instructions, and even if a human reviewer is “in the loop” somewhere they might never know that political instructions were built in. AI is so good at coming up with plausible-sounding explanations that a FOIA bot tuned to a particular set of goals could generate dozens of pages purporting to justify any redaction, even if it is contrary to black-letter law. Deployed at scale, it could cloak instructions by burying them in output.

Obscuring these types of instructions goes hand in hand with another major difference AI would bring: “automation bias.” This is the tendency of humans to “defer to automated systems,” leading to a reduction in “the amount of independent scrutiny that they exhibit when making decisions.” This is a well-documented phenomenon because it turns out that “people are bad at judging the quality of algorithmic outputs and determining whether and how to override those outputs.” If AI were to displace some or most of the FOIA processing pipeline, even humans somewhere in the loop might simply defer to whatever the AI says. Studies are already starting to show that using AI can reduce critical thinking, a finding that recalls Lisanne Bainbridge’s observations about the challenges of overseeing mechanized industrial processes in Ironies of Automation. We should be concerned that such tendencies, over time, will consolidate the locus of FOIA control higher up within the government.

Mitigating these risks will require extending current structural protections in FOIA. Congress wasn’t blind to the fox guarding the henhouse problem in creating this law, so review of the agency’s initial decision is already baked in, first through an administrative appeal and then de novo judicial review. The meaningfulness of these procedures must be preserved, both by disclosing additional information about AI’s usage and by limiting when it can be deployed.

My first set of risk mitigation proposals concerns the executive branch. Any AI FOIA tool development should be conducted in the open and, ideally, in collaboration with non-governmental experts and representatives of the requester community. Agencies should also be required to conduct periodic audits of any AI FOIA system and publish the results with enough detail for meaningful external review. These measures should not be controversial—the Justice Department’s FOIA handbook is public, as is its written guidance to agencies. OMB already requires “high-impact” AI applications to be subject to a number of additional procedures for testing and feedback. The government should extend these to FOIA systems to mitigate systemic risks.

When it comes to processing individual requests, agencies need to go further. Specifically, they must:

• Inform requesters whenever AI is used to respond to their request;
• Inform requesters how AI was used to respond to their request (e.g., for searching, for de-duplication, for exemption application);
• If AI was used to locate records, inform requesters as to which government databases/record repositories were searched;
• Preserve all of the inputs and outputs from AI systems that pertain to requests—the AI “administrative record”—so they can be reviewed; and
• Provide, upon request, a manual review and renewed decision regarding any action of an AI system.

These steps will not eliminate the risks of AI in the FOIA process, but hopefully they will lay the groundwork for ensuring it does not have the last say.

The second set of proposals concerns litigation. Judicial review is far from perfect (judges tend to agree with agencies most of the time), but it is the backstop for all of FOIA’s accountability. Indeed, the most impactful records released over the years have tended to come out in litigation. Preserving that independent review is a red line, and where AI’s involvement within the government must end. A situation where an agency can only gesture at the AI and claim it’s tokens all the way down cannot form the factual basis for judicial review. Once a case commences, defendant-agencies must conduct a manual review of their decisions and limit their litigation posture to the results of that human review.

What AI can do to improve the FOIA process is a question that should be discussed, experimented with, and empirically measured. But we must have guardrails in place to protect the law’s purpose first. Properly assessing and mitigating risk in the FOIA pipeline is overdue.

“I’m delighted to welcome Vanita Gupta and Karen Kaiser to the board as the Knight Institute enters its second decade,” said Lee C. Bollinger, president emeritus (ex officio) of Columbia University and chair of the Knight Institute’s board of directors. “Over the past ten years, the Institute has made important contributions to the defense of freedom of speech and the press. Those freedoms now face renewed and significant challenges, and Karen and Vanita will strengthen the board at a critical moment for our work.” 

The Knight Institute was established by Columbia University and the John S. and James L. Knight Foundation in 2016 to defend the freedoms of speech and the press in the digital age through strategic litigation, research, policy advocacy and public education. The Institute promotes a system of free expression that is open and inclusive, broadens and elevates public discourse, and that fosters creativity, accountability, and effective self-government.

“Karen Kaiser has spent her career protecting the people who report the news, and Vanita Gupta has spent hers making sure the law protects everyone equally,” said Jameel Jaffer, the Knight Institute’s executive director. “They are talented lawyers and inspiring leaders, and I’m delighted that the Institute will have the benefit of their experience and judgment.” 

Vanita Gupta is director of the Center for Law and Public Trust and a distinguished scholar in residence at NYU School of Law. From 2021 to 2024, she served as U.S. associate attorney general, becoming the first civil rights lawyer to hold one of the Justice Department’s top three leadership positions. Gupta previously served as president and CEO of the Leadership Conference on Civil and Human Rights and led the Justice Department’s Civil Rights Division during the Obama administration. Earlier, she was deputy legal director and director of the Center for Justice at the ACLU and an attorney at the NAACP Legal Defense & Educational Fund, where she led the effort to overturn the wrongful drug convictions of 38 people in Tulia, Texas. 

Karen Kaiser is director and legal advisor at the Clooney Foundation for Justice, where she works on harnessing AI to expand access to justice for women and journalists at risk. She previously spent 12 years as senior vice president, general counsel, and chief compliance officer at the Associated Press, where she oversaw the news organization’s global legal operations and worked on issues including press freedom, journalist safety, and AI governance. Kaiser was instrumental in reforms to U.S. Department of Justice guidelines governing subpoenas of members of the press in 2014 and 2022 and served on the DOJ News Media Task Force for more than a decade. She is chair of the International Bar Association Media Law Committee and serves on the steering committee of the Reporters Committee for Freedom of the Press. 

In addition to Bollinger, Gupta, and Kaiser, the Institute’s board includes Maribel Pérez Wadsworth, Knight Foundation President & CEO (ex officio); Daniel Abebe, Dean and Lucy G. Moses Professor of Law at Columbia Law School; Jelani Cobb, Dean and Henry R. Luce Professor of Journalism, Columbia Journalism School; Nicholas Lemann, Dean Emeritus at Columbia Graduate School of Journalism; Gerald Rosberg, former Senior Executive Vice President at Columbia; and Deb Roy, Professor of Media Arts and Sciences at MIT. 

For more information, contact: Lorraine Kenny, lorraine.kenny@knightcolumbia.org. 

Providing access and support for transparency is essential to civic engagement, self-governance, and protections against government corruption. It is the primary avenue for citizens to see what their government is up to and whether it is working in the public’s interest. However, whether by design or consequence, the federal FOIA process is nearly stagnant.

What we know:

• The number of FOIA requests is growing at an exponential rate.
• Federal FOIA staffing was heavily reduced at the hand of DOGE, and those unfilled positions mean fewer requests can be processed.
• Backlogs continue to be on the rise, as they have been for years, but they are now at their highest levels.
• FOIA professionals do not have adequate resources, specifically in staffing and technology, to resolve the backlog of requests at the federal level.
• Technology-assisted requests are making it easier for requesters to file extensive requests faster than humanly possible.
• Even if FOIA professionals want to make internal changes to increase efficiencies, they face obstacles within their agencies, slowing and blocking changes.

Access professionals are the front-line stewards of public access to government information. If they don’t have the time and tools to get their work done, information will not be publicly available. They are responsible for facilitating public access by receiving requests, tracking them, interpreting what is being requested, determining what will be released, levying fees when allowable, and resolving disputes with stakeholders. They must also ensure legal compliance through determining redactions, exemptions, denials, timelines, records management, and potential litigation. They do all of this while managing internal processing workflows—conducting records searches, reviewing and preparing records for release, keeping records organized and searchable, and following retention schedules.

It is a balance of legalities and logistics. Like any engine, it can’t run well without fuel, and in this context, that fuel is money, manpower, and technology. Agencies across the government are disparate in how they functionally respond to requests. Some are centralized, and some are decentralized. Some use advanced technology for search, review, redaction, and analytics; while others rely on spreadsheets and electronic filing. The needs of larger agencies that receive more than 100,000 requests per year would not compare to small agencies that receive fewer than 50 requests per year. But across most agencies, the backlogs are becoming too long to clear, and access professionals are just watching new requests join the queue with no hope of response any time soon.

A solution will require a multi-faceted approach to augment resources, both within agencies and through external funding structures: increase funding, analyze internal support operations, and review efficient technology use.

Increase funding

Getting the proper resources to drive more efficient processing takes money, and, at the federal level, that requires Congress to require agencies to properly fund the FOIA system. The FOIA Advisory Committee recently recommended that Congress provide funding directions that require agencies to consider their workforce size, the changes in requests in the past five years in volume and complexity, backlogs, and other operational structures (Recommendation 2026-03). In 2021, the Office of Government Information Services recommended that Congress ask the Government Accountability Office to study whether agencies are adequately funded and what they could accomplish with additional funding, specifically timely response and backlog reduction.

However, in this current administrative climate, it is unlikely that Congress as a whole will respond in a fashion supportive to FOIA. But individual lawmakers can make such requests of the GAO, or the comptroller general can initiate such studies. A potential catalyst would be (like the intent of FOIA itself) to shine light on the financial challenges agencies are experiencing.

Analyze operations

Agencies need to realistically analyze staffing levels and workloads to understand why the backlogs are increasing and what would be required to eliminate them. This analysis should also consider revising operational procedures. For example, do agencies need dedicated FOI professionals to respond to requests? Do they need centralized systems to respond? Do they need additional subject matter experts on staff to process complex requests?

In focus group research, FOIA professionals said that their work is rarely considered mission critical, and therefore they have trouble getting the proper resources they need to complete their work. As a result, the FOIA Advisory Committee recommended that agencies study the infrastructure needed to carry out their open government responsibilities (Recommendation 2026-11). Healthy FOIA operations run on proper staffing levels as well as appropriate technology to support productivity.

Review technology

Agencies should also review how they can adopt new technology to use for more efficient processing. Externally, the Office of Information Policy, the Office of Governmental Information Services, and the Department of the Interior are developing a FOIA technology inventory to evaluate what is currently in use. However, agencies themselves should study how technology can augment their own internal processes and how those processes can be modernized.

Adopting technology in government agencies is a long and protracted process. It is so difficult that technology can be outdated by the time agencies actually receive it. Technology adoption and acquisition is a slow and protracted process at the government level, and there have been previous calls for a revision of the process, especially for agency response to information requests. Additionally, there is no consistent system across agencies because agencies adopt the technology that makes the most sense for their needs and that fits their own budgets. This lack of interoperable technology across agencies is not a new obstacle, but it is becoming more apparent as other FOIA resources are shrinking.

Agencies’ piecemeal responses to the inability to respond to requests are not a long-term solution. For example, providing time extensions, a common response to the lack of capability to respond to requests, doesn’t solve the problem and disproportionately harms requesters. Agencies are simply moving the problem later in time rather than managing what is causing the problem. The requests are not the problem, the response capability is.

Charging more fees for voluminous or complex requests also ignores the root causes and unduly punishes requesters, especially when they are used as a cudgel to keep requesters at bay. Taxpayers already foot the bill for the FOIA process. Additional fees will not benefit the agencies charging them. Where fees are employed, they should be specifically structured to charge only what is necessary, leaving less room for weaponization.

Traditional solutions have not worked, and a concerted effort needs to be engaged. Fueling the response process with proper resources, staffing, and technology, all of which will need financial support, is essential. Unfortunately, that kind of tangible support is receding, and it is showing in backlogs and access professionals’ workloads. Ultimately, this failure will result in a lack of civic engagement and governmental accountability, which inevitably leads to eroded public trust and a weakened democracy.

Those legacies—and what it will take to confront them—were the focus of “The Forever War and the First Amendment,” a two-day symposium held at the Bollinger Forum on Columbia’s Manhattanville campus earlier this month that brought together journalists, scholars, lawyers, and activists to examine how the “War on Terror” transformed government power, the courts, the press, and democratic culture.

Co-hosted by Columbia’s Knight First Amendment Institute and the Simon and June Li Center for Global Journalism at Columbia Journalism School, in partnership with Harper’s Magazine and the Columbia Daily Spectator, the symposium examined government powers, legal doctrines, surveillance systems, and media practices that were in place before 9/11 and expanded afterward.

Opening the symposium, Jameel Jaffer, Knight Institute executive director, recalled the “apocalyptic feel” of New York City in the weeks after 9/11: smoke and ash in the air, fighter jets flying low over the city, soldiers in the streets, missing-person posters on every lamppost, and everyone on edge about the possibility of another attack.

“Many of us also had a sense of foreboding about how our government would respond,” Jaffer said. “Even in that moment of intense fear and sadness and confusion, it was clear that collectively we faced a momentous choice and that what our attackers did next was probably going to be less important, less consequential than what we ourselves did.”

What came next—the decisions, policies, and normalized ways of thinking that took hold in the years after—was the focus of the symposium.  

The Deeper Roots of the Forever War 

In her remarks, journalist Rozina Ali noted that she had expected her recently published book on Islamophobia, Seasons of Fury, would begin with 9/11, but her research led her to surveillance and law-enforcement practices going back to the 1960s. Artist and filmmaker Assia Boundaoui uncovered another example of pre-9/11 surveillance while making the award-winning documentary The Feeling of Being Watched and developing Inverse Surveillance Project: tens of thousands of pages of FBI files revealing that her Chicago-area Muslim community had been the target of a sweeping pre-9/11 counterterrorism investigation code-named “Operation Vulgar Betrayal.” 

“The data that is collected in the surveillance remains in perpetuity,” Boundaoui said. “There’s a double violence in this because it can be reused.” 

Across the symposium, speakers returned to a common pattern: Legal authorities and institutional practices that often predated 9/11 were expanded after the attacks and then normalized. They traced that development through the material-support laws, national-security surveillance, secrecy surrounding torture and targeted killing, the use of immigration powers against political dissent, and the weakening of institutions intended to hold executive power accountable.

When the Courts Deferred—or Did Not Decide 

ACLU attorney Brett Max Kaufman traced judicial deference to the climate immediately following the attacks: “Even judges that on 9/10 might have been inclined to be skeptical of government power or push back on certain kinds of arguments—on 9/12, that was no longer the case. That fear, the culture of censoriousness, and not knowing when the next attack was going to come affected them as well.” 

A day-two panel devoted to the First Amendment examined the consequences of that judicial deference. Panelists discussed cases ranging from Holder v. Humanitarian Law Project, in which the U.S. Supreme Court ruled that providing human rights training to a designated terrorist group could be a crime, to unresolved challenges involving warrantless surveillance and the punishment of noncitizens for political expression. Knight Institute litigation director Alex Abdo asked whether the courts had created “a jurisprudence of ignorance,” in which rights become “effectively unenforceable” because an array of doctrines allows courts to avoid deciding on the merits. 

Reporters for the Columbia Daily Spectator interviewed people whose stories intersected in important ways with the symposium’s themes. Alberto Mora, who served as general counsel of the U.S. Navy from 2001 to 2006, discussed his opposition to torture and the imperative to protect human dignity and the rule of law during a crisis. Mohamedou Ould Slahi, who was imprisoned without charge at Guantánamo Bay for 14 years and whose story was dramatized in the feature film The Mauritanian, described what happens when extraordinary powers created for supposedly dangerous people are turned against the innocent. 

How the Press Reinforced—and Exposed—the Forever War 

A panel moderated by Jelani Cobb, dean of Columbia’s Journalism School, examined how the Forever War both tested the press and exposed its weaknesses. Government surveillance, efforts to identify confidential sources, and secrecy surrounding the legal basis for torture made reporting more difficult. At the same time, news coverage often amplified official claims and helped legitimize government policies. Yet investigative journalism proved indispensable to exposing what officials had concealed. 

Dana Priest, whose exposure of the CIA black sites and other aspects of the counterterrorism campaign in The Washington Post earned her a Pulitzer, acknowledged that reporters and editors at both The Post and The New York Times failed to question official claims aggressively enough, while skeptical reporting was often given less prominence. 

A second-day panel examined how nonprofit newsrooms including Airwars, Drop Site News, and Zeteo are developing alternatives to legacy coverage, often through collaboration with local journalists and open-source investigators. Haley Willis, of Columbia Journalism School, described open-source reporting as “taking tools that kind of started in the hands of governments and militaries and turning them on them,” while cautioning that access to satellite imagery and other evidence can still be limited by cost and corporate or government pressure. 

The first day closed with a moving 45-minute performance by renowned pianist and composer Vijay Iyer, featuring work shaped by his response to 9/11 and its aftermath. Introducing the performance, Iyer described music as a way of reaching places words cannot, calling what followed “a kind of séance.” 

In connection with the symposium, the Knight Institute and Harper’s Magazine co-commissioned an essay by the writer Hari Kunzru about art created during this period. At the symposium, Kunzru joined journalists Steve Coll, M. Gessen, Nicole Hemmer, and Phil Klay for a panel about “The Forever War and Democratic Culture.” Kunzru’s essay in Harper’s is available online. 

Readers can watch recordings of individual panels of the Forever War symposium on the Knight Institute’s YouTube channel.

With the start of a new academic year upon us, as approximately 150,000 students are in their first weeks of law school, I’m writing a series of reflections on the conversations we had with podcast guests, thinking about the future of legal education and of professional responsibility.

For our first episode, Madhav and I talked with Kim Lane Scheppele, a leading expert on illiberal autocracy and democracies under stress. She described how the election of populist leaders, like Viktor Orbán in Hungary, is often followed by efforts to use the law to undermine constitutional institutions, and why this can draw lawyers into an authoritarian project even if they don’t intend to support it:

When you put into the process laws that actually have the effect of undermining democratic institutions, you get a lot more buy-in than if you simply did it by edict or by force, and so it legitimates the entire enterprise of undermining democratic institutions. That’s why law becomes so crucial. Now, who are the lawyers who do this stuff? Some of the lawyers who do this are lawyers who believe in the leader and believe in the program. But often times you get lawyers doing this because law is like a giant math problem if you like math.

It’s just fun to play with the rules and think, well gee, you know, how would you do this? How would you do that? Is there a way you can make an argument for this, that, and the other thing, and frankly, legal education in many places, and here I include the U.S., lends itself toward exactly this kind of lawyering. Because how often do we take first year law students and say, argue the opposite of what you believe…. So lawyers have learned to be hired guns. They’ve learned to think that clever legal arguments are good lawyering, and they have much less training and much less grounding in the values that underwrite democratic institutions, so that the lawyers themselves may not realize how much they're undermining democratic institutions with the very things they're recommending. (11:17-12:16, 13:07-13:32)

I’ve continued to think about Kim’s observation that many lawyers treat legal argument like they’re solving a math problem. There is nothing new, or inherently malicious, about lawyers making clever arguments in favor of their clients. Much of legal education is devoted to learning how to do this. Indeed, lawyers have an ethical duty to “zealously represent” their clients. However, lawyers are generally not trained to consider the potential consequences for democracy of their legal arguments—to ask themselves whether the legal rule they are advocating for may undermine the foundations of our democratic system. This is where at least some of our problems arise. And this makes me recall something else we discussed during our conversation with Kim: Many lawyers in the United States simply can’t believe our democracy could fail, that it is breakable.

This ties into the myth of American exceptionalism, of course. Most U.S. lawyers are educated in a system that inculcates the idea that we have a uniquely strong democracy, that our constitution is the best of all possible constitutions, and by extension, that though our system of government may be sorely tested, it will never truly come apart at the seams. If one believes that our democratic system will endure forever, come what may, then we’re all just playing a game to be the cleverest lawyer in the room.

Instilling in lawyers a sense of the ultimate fragility of democracy, and of their individual duty to protect it, is a project that will require the participation of many institutions, but law schools must play a fundamental role in it. Law students hear plenty about the constitution and at least a little about their duty to defend it. More rarely are they asked to seriously consider what duties they, personally, bear to uphold our democratic system, and most law students are given very little instruction about the conditions under which democracies erode or fail. I wonder if constitutional law classes would land differently if students learned that the United States only met the basic requirements for a democracy after the Voting Rights Act was passed in 1965. Or if every 1L took a capsule course on how Hungarian democracy was dismantled through legal maneuvering under Orbán. I hope you’ll listen to the full episode for more of Kim’s insights on the role of legal education in shaping lawyers who will stand up for democracy, and lessons from democratic backsliding in Europe.

The following can be attributed to Jameel Jaffer, executive director, Knight First Amendment Institute at Columbia University:

“We welcome Judge Kelly’s order. Trump’s ban harms both the press and the public, which relies on journalists to report independently on the president and his administration. We hope that Judge Kelly will soon make his temporary order permanent, and that his permanent order addresses the news organizations’ First Amendment claims, not just their Fifth Amendment ones. The First Amendment bars the president from transforming the press corps into a propaganda machine.”

The Knight Institute and RCFP’s amicus brief argues that the press has, for over a century, fulfilled a crucial role by reporting on newsworthy events from within the White House on behalf of the American people, and that the president’s decision to ban news organizations from the White House based on the viewpoints expressed in their reporting violates the First Amendment. Read the brief here.

Read today’s temporary restraining order here.

Read more about the case, CNN v. Trump, here. 

The Knight Institute has supported the Associated Press in its challenge to its exclusion from the White House, filing three separate amicus briefs over a period of 18 months. The Institute has also supported The New York Times in its challenge to restrictions on access to the Pentagon. Earlier, the Institute successfully sued President Trump over his blocking of critics from his Twitter account, which the Institute argued was a “public forum” for First Amendment purposes.

For more information, contact: Lorraine Kenny, lorraine.kenny@knightcolumbia.org. 

Of course, we need both stances. It’s hard to build something new when you’re getting run over and run off the field. Defense alone can’t produce a truly independent and in-depth press, free from undue corporate and government influence. But if we stop defending against blatant censorship and runaway consolidation, there’s little room left for viable alternative outlets.

The U.S. populace may not hold the media in particularly high regard at the moment. We face a fractured and truth-challenged world, with Trump poisoning public discourse and launching a decade’s worth of attacks on the press as “an enemy of the American people.” Yet, even as this authoritarian president attacks specific journalists, he retains the fig leaf of providing his followers with the “truth” on social media and decries only what he labels “fake news.”

So, if people agree that they want reliable facts and news—even if they can’t agree on what counts as a fact anymore—why do we rely on failing, advertising-driven business models and continually under-invest in the production of quality journalism and civic information? What problems stand in the way of producing this public good?

Trump himself, and his entire administration, are obviously a searing threat to journalism and to individual journalists’ livelihoods, safety, and freedom. Threats to shut out, silence, assault, or jail journalists are hardly a new or uniquely American phenomenon, but the rapid crescendo of such attacks under the second Trump administration is beyond alarming. At Free Press, we work to educate people about such harms, pass policies that protect against them (like shield laws and jawboning protections), and defend individual journalists from arrest and attack.

Free Press also fights corporate media consolidation, even though we don’t believe corporate media alone will produce journalism that serves all of us. There’s a tight nexus between authoritarianism and corporate consolidation. Merger parties will capitulate quickly to a corrupt regime in order to get their transactions approved. (It’d be nice to think that the fourth estate’s ethics make it immune to such tactics, but in an era of the Ellison family promising favorable coverage of the administration in exchange for antitrust approvals, we can’t fool ourselves.)

Recent developments like Disney standing up against the FCC’s threats to its ABC broadcast licenses, and even going on offense about the chilling impact of that agency’s actions, show that the tide may be turning. Big companies may have learned to stop routinely retreating, capitulating, and complying in advance when this administration pushes too far.

Yet even if the largest and most well-heeled outlets can stand up to those threats, there are still reasons to be concerned about media consolidation in authoritarian times.

For one, having fewer competing viewpoints and competing journalists is worse for expanding our horizons and uncovering the truth. For two, having fewer strong voices in the marketplace for ideas gives a government bent on censorship and narrative control a smaller number of chokepoints it needs to squeeze and coerce.

What’s more, even without an image-obsessed autocratic president overtly demanding more favorable coverage of himself and his administration, corporate consolidation is bad news. At Free Press, we’ve documented when fighting mergers and regulatory rollbacks that antitrust laws and broadcast ownership limits are necessary and effective in preserving local journalism, competition, and diversity.

The promise that larger conglomerates will take their excess profits and pour them back into improved and increased coverage is a phony, trickle-down economics argument. Increased hours of news after a television merger often means nothing more than duplicate stories and repeat newscasts, choked with watered-down national stories instead of truly local reporting. Companies insulating themselves from competitive pressures through mergers and acquisitions of rivals won’t invest more, they’ll just use the profits to pay shareholders or pay down the massive debts they ran up to make the merger happen in the first place.

Broadcast journalism and media is special, insofar as rules limiting ownership, reach, and even content have generally survived constitutional scrutiny till now. The “public interest standard,” based on a spectrum scarcity rationale, has meant that the FCC can to some degree decide who speaks over the nation’s airwaves. Brendan Carr has become Trump’s censorship czar, and whether cynically or fervently or a toxic mixture of both, he’s bent that standard perhaps beyond the breaking point in less than two years of Trump’s second term.

It’s even more difficult, in American jurisprudence and under the First Amendment, to do anything about the size and reach of corporate media companies that don’t use broadcast licenses. Free Press has used antitrust tools to beat media and tech mergers, including but not limited to broadcast deals. A storied line of U.S. Supreme Court cases confirms that undue economic concentration can violate federal antitrust law, even if the commercial “product” is protected by freedom of speech and freedom of the press. And there have been antitrust consent decrees and limits on non-broadcast industry sub-sectors, like film studio control of movie theaters and television network control of TV show production. But calls to strengthen or change our antitrust laws, to account not only for consumer welfare and workers’ rights but also political economy and even news concentration, haven’t resulted in any statutory changes yet. And those changes would be tested politically and constitutionally if and when they’re proposed.

In other words, advocates have tools to prevent undue concentration, but there’s no way to prevent all mergers and set a strict limit on reach. And even if there were, many First Amendment advocates would argue strenuously against such limits, as a cure that is worse than the disease of corporate consolidation—or at least dangerous in different ways.

Because we can’t limit the size and reach of all corporate journalism companies, and we recognize the pitfalls and dangers in trying to do so, Free Press has never been solely focused on striving for a more perfect corporate media. We need a more complete view of the journalistic ecosystem to compensate for the limits of corporate owners, publishers, and speakers.

We’ve written about how journalism is a public good that will be under-produced if it depends solely on an attention economy and sensationalist appeal, especially when the number of online content choices from human creators (and now AI too) is astonishingly large. That puts pressure on traditional cross-subsidies for news by advertising-supported television and print companies. And that prompts harmful calls to move money away from extractive tech platforms right back to the hedge funds and corporate conglomerates running the country’s largest newspapers and TV station groups. Free Press Action opposes these kinds of bills at the federal and state level too.

It’s also readily apparent that “if it bleeds, it leads” local news was never healthiest for the local communities doing the bleeding, and that a focus on the crime beat and other fear-mongering to get eyeballs on the nightly newscast perpetuates racism but doesn’t produce the kind of civic information people need.

That’s why Free Press has always seen public media as a key complement to corporate outlets. Communities benefit from having a wide range of community-centered, non-commercial, and non-profit newsrooms, with at least the potential for government funding. Public money is not anathema to a free press, so long as there’s a firewall between the legislative decisions on such funding and the editorial decisions made by its recipients. An ecosystem that includes robust public media improves the quality and amount of accountability journalism produced.

It’s no accident that an authoritarian Trump administration finished the job started decades ago by Republican lawmakers to zero out federal funding for public broadcasting. Free Press fought that rescission. More than PBS, NPR, and the largest non-commercial stations, those cuts hurt stations serving less densely populated, less privileged, and less affluent communities, as well as independent filmmakers and documentarians whose work finds a home on public media. Yet the prior Corporation for Public Broadcasting was never enough. We need to reimagine public media, to make it more decentralized, truly local, and resilient to political funding decisions.

That doesn’t mean taking government funding out of the equation, and forcing non-commercial stations to rely solely on “underwriting” or charity from well-to-do donors. It means projects like the New Jersey Civic Info Consortium that Free Press helped conceive and create. These kinds of models are gaining momentum in other states too. The idea is to gather public money from general treasury revenues or dedicated tax and funding streams, then administer newsroom grants independently to foster the production of civic information and responsive journalism.

There’s no silver bullet or single answer to preserving quality journalism against the dangers of corporate consolidation and government censorship. But public dollars drawn from diverse sources and administered independently can fund non-commercial journalism by and for the people in communities never fully served by corporate media outlets.

One of the most common refrains about the federal Freedom of Information Act (“FOIA”) within the requesting community is that the statute feels broken. Even before the changes that have occurred under the current administration, agency FOIA programs were understaffed and underfunded. There has been a trend toward less disclosure and greater secrecy that transcends party lines. And apart from abuse of exemptions, agencies have long sought to erect extra-statutory procedural obstacles like sensitive review and equities-based consultations, or creative schemes such as “still interested” letters and administrative closure.

Congress has not been oblivious to these developments. The FOIA Improvement Act of 2016, for example, was designed to ameliorate many of the problems requesters face today. The Act codified a foreseeable-harm standard and twenty-five year “sunset” rule for the deliberative-process privilege. It introduced a “rule of three” to require affirmative disclosure of frequently requested records in online “reading rooms.” Fee limitations based on the failure to follow the statute’s deadlines were intended to incentivize faster processing. Yet, despite the merit of those changes, they have proven inadequate. A decade later, in the face of a record-breaking number of incoming requests, FOIA hardly seems fixed.

For some requesters, litigation is the only way out of this mess. Congress tasked the courts with enforcing agencies’ disclosure obligations. But, in practice, judicial review falls short of Congress’s design. As I have explained elsewhere, in a proposal for a dedicated Article III “FOIA Court,” “the volume of FOIA litigation nationwide has outpaced the ability of federal courts to deal with those cases in a timely manner.” FOIA suits often fail to receive the attention they deserve. More troublingly, a certain functionalism seems to prevail among jurists. Judges often resolve cases based primarily on uncritical review of agency declarations, with limited adversarial testing. Judicial review is plagued by policy concerns, information gaps, and doctrinal ambiguity.

We need to restore the independent judicial role that FOIA promises to the public. FOIA’s shortcomings are not the result of formal Chevron-style deference. The statute expressly prescribes de novo review and, for the most part, judges review legal questions in that manner. FOIA’s judicial-review problem instead arises from a functional deference driven by jurisdictional and remedial uncertainty, the anomalous procedural architecture of FOIA litigation, and atextual judicial doctrines. Congress needs to address these failures together, through both corrective clarifications of the statutory text and substantive reforms that improve the litigation process.

Rigorous judicial review would not only provide a robust check on agencies but also offer relief to aggrieved requesters. It would positively influence FOIA processing more broadly, too. Although only a small number of requests end up in court, caselaw developments and litigation practices heavily influence FOIA processing at the administrative level. Everyone benefits when courts are doing their job.

Clarifying Subject-Matter Jurisdiction

As the Supreme Court has explained, “without jurisdiction a court cannot proceed at all in any cause.” FOIA, for instance, provides that a federal court “has jurisdiction to enjoin [an] agency from withholding agency records.” Frustratingly, some agencies argue this language implicates the court’s authority to hear a case in the first instance. But that interpretation improperly transforms the elements of a FOIA claim—an allegation “an agency has (1) ‘improperly’; (2) ‘withheld’; (3) ‘agency records’”—into a set of jurisdictional prerequisites.

This confused position, if widely adopted, would have serious implications. At the least, it requires a requester to prove the merits of a claim at the pleading stage, thus reversing the burden placed on an agency to sustain its action. That is problematic for two reasons. First, because it contradicts the well-established principle that “[t]he burden is on the agency to demonstrate, not the requester to disprove,” that the records at issue are not subject to FOIA or have not been improperly withheld. And second, because most requesters, as a practical matter, have no access to the facts needed to prove or disprove the elements of either their claim or the agency’s affirmative defenses. The government holds all the cards. (We will return to this problem of “informational asymmetry” in a moment.)

Although three circuits have rejected this jurisdictional argument, the Department of Justice continues to advance it to the detriment of requesters. Providing clarity on jurisdiction—and putting an end to the government’s game of shifting evidentiary burdens to requesters at the complaint stage—is essential for making sure FOIA litigation functions as Congress intended.

Explicating Remedial Authority

A right of access without an effective remedy for noncompliance is meaningless. Although FOIA authorizes the courts to “order . . . production” and “enjoin . . . withholding[s],” contemporary cases frequently raise broader remedial questions, which are construed as constructive withholdings to fit within this textual scheme. This is particularly true with the allegations underlying so-called policy-or-practice claims, which depend on the notion that “FOIA imposes no limits on courts’ equitable powers in enforcing its terms.” It is also true with claims involving FOIA’s affirmative-disclosure requirements.

As to the latter, courts generally agree they can hear claims over an agency’s failure to publish records in the Federal Register or post them in an online “reading room.” But at least one court of appeals has determined it cannot provide common-sense relief after concluding an agency has neglected its affirmative-disclosure obligations. That is, instead of entering an injunctive order that requires an agency to comply with its reading-room duties, the U.S. Court of Appeals for the District of Columbia Circuit has stated the available relief is limited to enjoining the agency, even prospectively, to producing records merely to the prevailing plaintiff. The argument for limiting the remedy in this way is that a more general publication order would go beyond curing an individual plaintiff’s informational injury. Two other circuits have found that position unpersuasive. 

At least one circuit judge, in concurrence, has cast doubt on the validity of policy-or-practice claims altogether, thereby opening the door to further erosion of courts’ remedial authority and the diversity of claims the statute envisions. Admittedly, the policy-or-practice theory is an odd outgrowth of a jurisprudential mootness exception that finds little purchase in the text of FOIA itself. But the theory is well-established and, more than anything, this shortcoming counsels in favor of Congress codifying such a claim to prevent its outright rejection by the courts. 

The government has also sought to leverage remedial authority against requesters, arguing that FOIA supplements a court’s inherent equitable power to enter “clawback” orders forcing requesters to return or destroy public records that an agency maintains were inadvertently disclosed. The D.C. Circuit has largely rejected the availability of “clawback”—albeit without addressing lurking First Amendment objections—yet the government is attempting to force the issue again.

Ultimately, Congress must carefully spell out a court’s remedial authority and the available causes of action a requester may bring under FOIA. By providing this textual clarity, Congress would reduce collateral disputes while reemphasizing that courts are meant to function as the “enforcement arm of the statute,” fashioning such relief as needed to fulfill the law’s promise of openness and transparency.

Restoring Meaningful De Novo Review

FOIA’s de novo standard is one of its most important features. Congress understood that agencies have institutional incentives to withhold embarrassing, inconvenient, or sensitive information. Hence the need for independent review. Unfortunately, many courts are not holding agencies to account. They have accepted agency efforts to “circumvent[]” de novo review by engaging in various types of “deference conservation,” especially in cases implicating national security or law enforcement matters.

This “deference” extends beyond legal questions—or the meaning of statutory terms, including the scope of FOIA’s exemptions—to factual matters or the application of law to facts. Yet these are the situations in which a fresh look by a judge would be most appropriate under FOIA. In one older district court case, Chemical Manufacturers Ass’n v. CPSC—which the government continues to cite in contemporary FOIA litigation—a judge decided he owed “considerable deference to the [agency’s] judgment as to what constitutes . . . ‘part of the agency give-and-take’—of the deliberative process,” at least so long as decision-making was still ongoing. That kind of solicitude for an agency’s evaluation of “what confidentiality is needed”—a kind of epistemic deference that relies on an agency’s supposed expertise—seemingly veers quite far from de novo review.

As others have explained, deviation from the de novo standard is also found in a court’s uncritical acceptance of declarations and Vaughn indices, as well as its reliance on substantive or procedural interpretations of FOIA that are structurally oriented toward affirming the government’s position. This problem of procedural deference extends to judicial doctrines like the presumptions of “good faith” and “regularity,” and the D.C. Circuit’s more controversial presumption of compliance with FOIA’s duty to segregate non-exempt material from otherwise exempt records.

Although informal deference—including the practice of allowing agencies repeated opportunities to supplement the evidentiary record with multiple rounds of summary judgment—may seem reasonable to some, disregard for the standard set by Congress has transformed much of FOIA litigation into a sort of box-checking routine in which attorneys and judges “focus[] on a dispute about the adequacy of the [agency’s] Vaughn index, rather than a dispute about the merits of the exemption claims themselves.” It is also unclear whether the practical erosion of de novo review can persist within the new “Loper Bright interpretive paradigm” of administrative law, with its emphasis on the role of independent judicial judgment. In any event, Congress needs to reinforce and empower judges to independently verify an agency’s factual claims whenever a requester raises genuine questions about search adequacy, exemption use, segregability, delay, or systemic noncompliance with the statute.

Accounting for Information Asymmetry

As the D.C. Circuit has explained, the “unique evidentiary configuration” of “the typical FOIA case ‘distorts the traditional adversarial nature of our legal system’s form of dispute resolution.’” Requesters face an “asymmetrical distribution of knowledge” in challenging agency action because “the agency alone possesses, reviews, discloses, and withholds the subject matter of the request.” Yet that “nearly impregnable defensive position” depends on the agency’s ability to meet its burden “to sustain its action.” But a nominal burden on the agency is not enough. Basic principles of fairness require agency arguments to be subject to genuine adversarial testing. The unavailability of discovery in most cases, total absence of an administrative record, and selective application of the federal rules of procedure and evidence frustrate an equitable playing field.

Congress should adopt procedures that reduce this imbalance. It could require agencies to proffer more detailed declarations, strengthen Vaughn index requirements, authorize discovery when there is a material factual dispute, encourage the use of FOIA trials, reinforce in camera review, and make clear that agencies bear a meaningful burden to justify the reasonableness of their searches, withholdings, foreseeable-harm determinations, and segregability efforts.

Conclusion

FOIA is often considered an administrative problem. Agencies lack resources, backlogs continue to grow, and requesters face interminable delays. Those concerns are real, but they do not fully account for the statute’s declining effectiveness. At its core, FOIA enshrines a judicially enforceable right of access. Congress vested federal courts with responsibility for independently evaluating agency action under FOIA. Decades of litigation and institutional cultural shifts have eroded that design. The result is a system in which courts too often validate agency secrecy rather than engaging in true independent judgment with an eye toward openness. Congress should act to restore the judiciary to the role it envisioned.

The amicus brief argues that the press has, for over a century, fulfilled a crucial role by reporting on newsworthy events from within the White House on behalf of the American people, and that the president's decision to ban news organizations from the White House based on the viewpoints expressed in their reporting violates the First Amendment.

On September 24, 2026, the district court granted the plaintiffs' motion for a temporary restraining order, requiring the government to restore the plaintiffs' access to the White House.

Status: On October 13, 2026, the district court granted the plaintiffs' motion for a preliminary injunction, requiring the government to maintain the plaintiffs' access to the White House while the case proceeds.

Case Information: Cable News Network, Inc. v. Trump, No. 1:26-cv-03287 (D.D.C.).

But it is not just requesters who register high levels of frustration. As the number of requests increases almost exponentially year over year and the volume of records created and maintained by agencies balloons, the resources agencies are given—both human and technological—perennially fall far short of what is needed. On one side, FOIA offices struggle to communicate with requesters effectively even when communication would be mutually beneficial. And on the other side, those same offices face barriers in securing the cooperation of agency program officials in other units to search for and produce responsive records. Agency officials cite the near impossibility of the task at hand, and having been set up to fail, FOIA compliance contributes to a public perception of government incompetence.

In an attempt to support agencies in their FOIA processing, tracking, referring, communicating, and more, private companies have long offered various purported solutions. The leading product, FOIAXpress, has been used by many agencies for more than two decades, but competitors have cropped up like NextRequest and Armedia. Without doubt, these private companies are innovating in the FOIA processing space, including by trying to integrate across non-FOIA staff and by harnessing the power of artificial intelligence. Still, though, private products are adopted piecemeal, agency-by-agency, and do not integrate across the government. Moreover, the nature of private products has exposed government to potential security breaches. And finally, these products are expensive and offer many add-ons and customization options, making even the very same products function differently across agencies. For requesters, this means that each agency still has a bespoke requesting process.

At times, both Congress and the administration have recognized the need for greater ownership and consistency of technology in a centralized location in government. In the FOIA Improvement Act of 2016, Congress required the Office of Management and Budget to “ensure the operation of a consolidated online request portal that allows a member of the public to submit a request for records… to any agency from a single website.”

While this had the promise of taking a step in the direction of improving FOIA infrastructure, the implementation has proven wholly inadequate. FOIA.gov, the government’s attempt at moving toward a FOIA portal, acts mainly as a redirect site; that is, it provides a centralized area to find a link to each agency’s request submissions process. It does not, itself, allow a single submission site.

This is not the first attempt that has gone awry. The EPA spearheaded an earlier effort, FOIAonline.gov, a portal system that allowed other agencies to opt in to using it for request submissions on a single site and was used by twenty-two agencies at its height. But after the EPA experienced high costs in maintaining the system, it was decommissioned and shut down entirely in 2023.

Even at their zenith, none of these efforts have been adequately ambitious to produce the kinds of gains that real technology infrastructure might offer. Other countries have vastly outpaced us in this regard, and their models are instructive in the possible efficiency gains of technology platforms that serve the administrative needs of FOI professionals and the public alike.

During the height of Mexico’s transparency regime, Mexico’s then-extant Information Commission created a National Transparency Platform, an integrated portal system that out-imagines anything we have tried. As a basic matter, it allowed requesters to make a request to any governmental body at any level of government from a single portal, or even to make a request to multiple agencies at once. It also allowed the requester to log in and see all past requests made, with deadlines for responses clearly indicated, as well as all subsequent communication and any eventually released records. The portal served as a repository for all released records which (absent individual privacy concerns) were also posted publicly for anyone to search and review. All records required to be made affirmatively available were likewise housed on the portal.

But where it really shone was on the backend. The portal served as the internal processing management system for government entities. Each agency unit that processed requests (the equivalent of a FOIA office) could see all the pending requests and deadlines, could refer the request internally to the program or component of the agency that was likely to have responsive records—each of which had a designated contact for handling FOI requests—and could set an internal deadline for that program to respond to the FOIA office. The contact point in the program office, likewise, had their own login that provided a list of all referred requests pending and allowed them to build the record on their end.

Finally, the National Transparency Portal also facilitated communication between the agencies and the information commission; once the agency response was registered, it allowed requesters, with the click of a button and a sentence-or-two explanation, to administratively appeal the response to the commission. The commission then had access to the full administrative record of what happened prior to the appeal and could receive additional documentation through the portal, as well as request additional information from the agencies. At base, therefore, the portal looked different to each actor who had a touchpoint on a request by virtue of their login. It clearly managed that individuals’ workflow and facilitated retention of a full record of what happened.

Mexico is far from the only country to have developed such a deceptively simple sounding system, but one that radically transforms the FOI space. Brazil, for example, has a very similar one-stop-shopping portal system for their FOI space, as do Paraguay and Norway.

The U.S. needs a true FOIA portal. A real portal would create the infrastructure that would facilitate the possibility of something akin to full compliance with FOIA obligations. It would supplant a piecemeal system with a system designed for this purpose, built in house, maintained by government computer engineers, and consistent across agencies. It would streamline the workflow inside agencies, make clear who has the responsibility for the request at each stage, and create priorities for work based on deadlines. It would provide the public with one-stop-shopping for government information and lower the barriers to use by ordinary citizens. And it could create efficiencies for agencies, which would no longer need to invest FOIA budgets in procurement of third-party products for FOIA solutions.

The most exciting part of a true portal, however, is the nearly endless list of possibilities beyond improving the current request-and-response compliance rate. For example, a portal could facilitate release-to-one-release-to-all, a widely celebrated concept that once a record is released under FOIA it should be available publicly for anyone to access. The DOJ has supported the idea in pilot version previously, but it never became a reality. Indeed, implementing a release-to-one-release-to-all system would obviate the need for FOIA’s current requirement that agencies release publicly any record requested three or more times, a requirement agencies have found extremely difficult to implement. In fact, much more broadly, a portal creates an opportunity for meaningful affirmative disclosure regimes in which records are searchable government-wide and agencies have checklists of disclosure obligations with which they must comply.

Exemption application could be improved through the use of a portal as well. For example, a portal could collect the body of past agency decisions—both at the initial and the administrative appeal level—as well as adjudicated case law in one place. It also would contain all previously released records—whether at the administrative level or in response to litigation—thereby making plain when information is already public on a particular topic and facilitating a determination that further release would not harm a protected interest. Easy access to more definitive answers on the application of exemptions or other matters under the law might improve agency compliance in future cases.

A portal might also be used to facilitate data analysis, auditing, and other programmatic evaluation of FOIA offices. Patterns of violations—whether delay, overwithholding, or failure to engage in affirmative disclosure—could benefit from systematic solutions, and a portal could be used to facilitate identification of areas ripe for improvement.

Even more broadly, a portal could facilitate many of the other types of FOIA reforms that have been considered within the FOIA community. For example, it might allow tracking the personal responsibility attached to FOIA violations to particular employees in a way that could sanction flagrant noncompliance. It could facilitate improved judicial review by creating a full administrative record for the court and permitting easier in camera review of records. It could link to declassification efforts by the National Declassification Center or be integrated as a tool hosted by the Archives into the NDC’s work. It could be integrated into artificial intelligence solutions; imagine, for example, an AI tool that combed through agency records and identified records that are likely of interest to the public and can be affirmatively disclosed. And it may hold the potential for many more uses we could tailor the portal to as it developed over time. FOIA needs a portal now.

“Journalists covering the Pentagon play a crucial role in keeping Americans informed about the actions that the U.S. military and its leaders take in our name,” said Jake Karr, staff attorney at the Knight Institute. “These new restrictions represent a transparent attempt to cow independent, fact-based reporting and convert the Pentagon press corps into something more like a propaganda machine.”

Today’s brief argues that the government’s unprecedented restrictions violate the First Amendment’s public forum doctrine because they are impermissibly motivated by a desire to suppress critical journalism and because they undermine the very purpose of allowing journalists to access the Pentagon in the first place—to facilitate accurate and timely reporting on matters of war and peace that informs and serves the public. It urges the court to roll back these restrictions and restore the Pentagon press corps’ ability to gather and report the news without fear or favor.

“The administration is trying to prevent journalists from publishing reports about the military that it simply doesn’t like,” said Katie Fallow, deputy litigation director at the Knight Institute. “Especially in times of war, the public needs accurate, timely, and independently sourced information that only a free press can provide.”

The Department of Defense issued the interim policy that is the focus of this appeal following a ruling by the district court that the original policy restricting reporters’ access to the Pentagon was unconstitutional; the government has not challenged that ruling.

The Knight Institute filed a similar amicus brief earlier this month in a related case also challenging the interim policy.

Read today’s amicus brief here. 

Read more about the cases—New York Times v. Department of Defense (NYT I and II)—here. 

Lawyers on the case include Katie Fallow, Jake Karr, Jameel Jaffer, and Alex Abdo for the Knight First Amendment Institute.

Today, the Knight Institute also announced that it will support CNN, MS NOW, and Politico in their lawsuit against the Trump administration’s White House press ban. Read more here.

For more information, contact: Lorraine Kenny, lorraine.kenny@knightcolumbia.org.